Identifying similarities at metabolic pathways with a strategy of Enzymatic Step Sequences

An easy and fast strategy to compare functionally the metabolic maps is described. The KEGG metabolic maps are transformed into linear Enzymatic Step Sequences (ESS) using the Breadth First Search (BFS) algorithm. To do this, the KGML files are retrieved, and directed graph representations are created; where the nodes represent enzymes or enzymatic complexes, and the edges represent a compound, that is the 'product' from one reaction and a 'substrate' for the next. Then, a set of initialization nodes are selected, and used as the root for the construction of the BFS tree. This tree is used as a guide to the construction of the ESS. From each leaf (terminal node), the path is traced backwards until it reaches the root metabolic map and with two or fewer neighbors in the graph. In a second step, the ESS are compared with a Dynamic Programing algorithm, considering an "ad hoc" substitution matrix, and minimizing the global score. The dissimilarity values between two EC numbers ranged from 0 to 1, where 0 indicates similar EC numbers, and 1 indicates different EC numbers. Finally, the alignment is evaluated by using the normalized entropy-based function, considering a threshold of ≤ 0.27 as significant.•The KEGG metabolic maps are transformed into linear Enzymatic Step Sequences (ESS) using the Breadth First Search (BFS) algorithm.•Nodes represent enzymes or enzymatic complexes, and the edges represent a compound, that is 'product' from one reaction and a 'substrate' for the next.•The ESS are compared with a Dynamic Programing algorithm, considering an "ad hoc" substitution matrix, and minimizing the global score.